On Sun, Jan 12, 2020 at 4:50 AM Gunnar Þór Magnússon <gun...@magnusson.io> wrote: > > > I thought that upgrading is not that simple. > > It depends. > > If you have Python 2 code that deals with a lot of text in byte form, and > it's kind of vague where you convert from bytes to strings, you may have a > bad time. >
It's always hard to port buggy code or underspecified code. If you have code written in some hypothetical language that doesn't distinguish properly between 16-bit ints and 64-bit floats, and has sloppy conversions between them and different semantics, then it would be a pain to port that to any other language. Ultimately, the cure is to figure out the programmer's original intention and implement that. > Otherwise, it may not be that bad. I ported around 500k lines of Python 2 to > 3 this year, and it went smoothly. The most valuable resource I found was > eevee's post on the subject: > Python <python-list@python.org> Indeed. Especially if you're porting to a fairly recent Py3, chances are that most of the code will work just fine. There'll be just a few places where you have to manually figure things out, and for the rest, automated conversions like 2to3 will cover it. (And a lot of it doesn't even need automated conversion. There aren't actually THAT many things to change.) ChrisA -- https://mail.python.org/mailman/listinfo/python-list
